A Critical Review of SERVE-HF Follow-Up Studies and Their Impact on Clinical Practice

Purpose of Review: The SERVE-HF trial results were widely interpreted as strong evidence that adaptive servo-ventilation was contraindicated for the treatment of central sleep apnea/Hunter-Cheyne-Stokes breathing in patients with heart failure and reduced ejection fraction. Subsequently, the investigators have published five follow-up analyses of the study data. This paper will review these publications and provide additional commentary meant to provide updated context concerning this important subject. Recent Findings: The investigators have analyzed their dataset using different approaches and extracted additional information. Inevitably, some of the original criticisms leveled at SERVE-HF continue to apply, but most add important new information concerning their endeavor. One of these studies provided an on-treatment analysis that supported the original findings, while another reinforced the likelihood that excess mortality in the active treatment arm was due to a cardio-electrical event during wakefulness. A third study concentrated on patients with the HCSB pattern of CSA and revealed a relationship between cycle length and the primary composite endpoint in subjects belonging to both the active treatment arms and the control group. This analysis is of interest for the study of heart failure in general and did not necessarily confirm or clarify the findings of the original study. Summary: The most important paradigm shift in the management of central sleep apnea in patients with systolic heart failure is attributable to SERVE-HF. The results of the trial contradicted all heretofore accepted evidence in the field including the assumptions of its investigators. While the primary outcome of the trial was negative, the main reported finding of this was driven by an exploratory analysis that was not part of the statistical original design of the trial and that could not be explained by any mechanistic analysis from the trial data.